Double Crosslinked Polymer Electrolyte by C–S–C Group and Metal–Organic Framework for Solid‐State Lithium Batteries

Poly(ethylene oxide) (PEO) is among the most promising candidates for solid‐state electrolytes in lithium metal batteries. However, the low ionic conductivity caused by strong coordination between Li ions and the EO chains limits the practical application of PEO‐based electrolytes. Herein, a double crosslinked PEO‐based electrolyte with alternate C–S–C groups and functionalized metal–organic frameworks (MOFs) is proposed. The incorporation of C–S–C groups not only accelerates Li ions transport by weakening the coordination between Li ions and polymer backbone, but also facilitates segmental relaxation of the polymer backbone. The PEO‐based electrolyte with C–S–C groups shows a remarkable 13‐fold increase in ionic conductivity. Furthermore, when functionalized MOFs are used as crosslinked centers, the double‐crosslinked PEO‐based electrolyte with a robust network structure possesses enhanced mechanical/electrochemical/thermal stability and limited anion transmission. As a result, the symmetrical Li||Li cell enables over 2400 h cycling at room temperature. The LiFePO4||Li cells show long cycle life over a wide temperature range from 25 to 100 °C, and a high areal capacity of 1.43 mAh cm−2 is achieved with a cathode loading of 10.0 mg cm−2. This study demonstrates a promising strategy to develop advanced electrolytes for potential solid‐state lithium‐metal batteries.

directly on the polymer backbone of PEO, polymer backbone design can construct a weak Li þ -polymer coordination environment. A class of main-chain fluorinated SPE enables fast Li þ transport and an extended electrochemical window benefitting from the weak coordination between Li þ and EO. [19] Besides, a loosely coordinating SPE with a low density of oxygen sites provides higher ion conductivity and Li þ -transference number than PEO. [10] However, the nonactive groups (F and C atoms) inserted in the EO chains extend the distance between Li þ coordination sites, which may impede Li þ transport on the polymer chains. [20] The C-S-C group is a kind of coordination group of Li þ with a coordination mechanism similar to that of the C-O-C (EO) group. [21] The binding energy between Li þ and the C-S-C chain is about 74% of that between Li þ and the C-O-C chain, contributing to faster Li þ transport in the C-S-C chain than that in the C-O-C chain. [22] However, the low binding energy between Li þ and the polymer chain will lead to adverse Li-salt dissociation and insufficient Li þ concentration. [22,23] In this regard, it is viable to incorporate C-S-C groups into the PEO backbone to weaken the binding energy between Li þ and the C-O-C chain. The C-S-C groups can weaken the coordination between Li þ and the polymer backbone to accelerate Li þ transport, while the C-O-C chain can facilitate the dissociation of Li-salt to ensure sufficient solvated Li þ . But the poor electrochemical stability of C-S bonds and poor mechanical/thermal stability of linear polymers limit their potential application in lithium metal batteries. [10,22] Metal-organic frameworks (MOFs), consisting of organic linkers and metal ion clusters, possess both the characteristics of organic and inorganic materials. The organic linkers of MOFs have the potential to be covalently linked with a polymer matrix to construct a crosslinked structure, and the MOFs as fillers in SPEs are believed to enhance the stability and tether anions. [24,25] In this work, a double crosslinked PEO-based electrolyte with C-S-C groups and functionalized MOFs (MIL-101-NH 2 ) was prepared via one-step photoinitiated "thiol-ene" click chemistry ( Figure 1a). The incorporation of C-S-C groups in the PEO backbone helps to weaken the coordination between Li þ and the polymer backbone, and accelerate segmental relaxation of the polymer backbone, which increases the ionic conductivity by 13 times. After cross-linking with functionalized MOFs, the double crosslinked PEO-based electrolyte exhibited an enhanced Li þ -transference number of 0.46 and improved mechanical/ electrochemical/thermal stability. By these advantages, the symmetrical Li||Li cell employing M-P(EO 20 -ES) could stably cycle for more than 2400 h at 0.05 mAh cm À2 and room temperature. The LiFePO 4 ||Li cell showed a capacity retention of 81% after 1000 cycles with a superior Coulombic efficiency (99.65%) at 50°C and demonstrated long cycle life at 25 and 100°C. More importantly, the LiFePO 4 ||Li cell with cathode-supported M-P(EO 20 -ES) showed a high areal capacity of 1.43 mAh cm À2 with a cathode loading of 10.0 mg cm À2 .

Results and Discussion
The incorporation of alternate C-S-C groups into the backbone of PEO was realized via thiol-ene click-polymerization, which is demonstrated to be a potential reaction to synthesize polythioethers. [26] Specifically, the terminal -SH groups of SH-PEG-SH reacted with the terminal C¼C groups of poly(ethylene glycol) diacrylate (PEGDA) and functionalized MOFs, to obtain a double crosslinked PEO-based electrolyte with C-S-C groups and functionalized MOFs (Figure 1a). The functionalized MOFs bearing polymerizable C¼C groups were synthesized by postsynthetic modification. [25] The functionalization of MOFs (MIL-101-NH 2 ) was confirmed by X-ray diffraction (XRD) and Fourier transform infrared spectroscopy (FTIR). The identical XRD pattern of the functionalized MOFs and pristine MOFs ( Figure S1, Supporting Information) suggests that the functionalized MOFs retain their crystal structure after postsynthetic modification. Meanwhile, the characteristic FTIR peak of C¼C at 1653 cm À1 appeared on the functionalized MOFs (Figure 1b), indicating that the functionalized MOFs can serve as crosslinked centers of polymers to construct a double network structure with covalent bonds.
The photopolymerization of precursors was confirmed by FTIR ( Figure S2, Supporting Information). The characteristic vibration peak of C¼C groups between 1615 and 1640 cm À1 is invisible in the as-prepared M-P(EO 20 -ES). [27] The complete reaction of the precursors was further corroborated by Raman spectra (Figure 1c), from which the peaks at 1637 and 2569 cm À1 (corresponding to the stretching vibrations of C¼C in PEGDA and S-H in SH-PEG-SH, respectively) disappeared completely after polymerization. At the same time, the stretching vibration peak of C-SH in SH-PEG-SH (667 cm À1 ) blue-shifts to the peak of C-S-C (671 cm À1 ). A similar shift was detected by X-ray photoelectron spectroscopy (XPS), as shown in Figure 1d. The S 2p peaks of SH-PEG-SH at 163.58 and 164.80 eV (assigned to the 2p 3/2 and 2p 1/2 peaks of C-SH) shift to 163.82 and 165.05 eV (assigned to the 2p 3/2 and 2p 1/2 peaks of C-S-C) in M-P(EO 20 -ES). [28] The Raman and XPS results confirmed the conversion of thiol groups and the formation of C-S-C groups in the backbone of M-P(EO 20 -ES).
To evaluate the physicochemical property of M-P(EO 20 -ES), P(EO 20 -ES) was prepared in the same way without functionalized MOFs, M-PEO 20 was prepared by the polymerization of functionalized MOFs and PEGDA, and PEO 20 was prepared by the polymerization of PEGDA for comparison. The XRD was conducted to investigate the phase of the polymer matrix ( Figure S3, Supporting Information). The broad peaks between 13°and 27°of the as-prepared electrolytes in the XRD patterns confirmed their amorphous feature, which is beneficial for Li þ transport. The transport of Li þ in amorphous polymers is strongly related to the polymer segment relaxation. [29] Therefore, the thermal behaviors of SPEs were studied by differential scanning calorimetry (DSC). As shown in Figure 2a, the as-prepared electrolytes exhibit broad melting peaks due to the disorderly accumulation of crosslinked polymers. Moreover, the glass transition temperature (T g ) of P(EO 20 -ES) (À50°C) and M-P(EO 20 -ES) (À47°C) is much lower than that of PEO 20 (À35°C) due to the long bond length and low internal rotation barrier of C-S bonds in the backbone, indicating more significant polymer segmental relaxation at room temperature. [30,31] The M-P(EO 20 -ES) also demonstrates good thermal stability due to the double crosslinked network structure, as evidenced by thermogravimetric analysis (TGA) in Figure 2b. The initial degradation temperature of M-P(EO 20 -ES) is 10°C higher than that of P(EO 20 -ES) and PEO 20 . In addition, a high-temperature tolerance test of M-P(EO 20 -ES) was presented and compared with a commercial polypropylene (PP) separator ( Figure S4, Supporting Information). The PP separator became crinkly after heating at 150°C for 2 h and melted after heating at 200°C for 2 h. In contrast, no obvious changes in dimension or morphology were observed in the M-P(EO 20 -ES) at 200°C, showing its potential application in the field of high-temperature lithium metal  batteries. [32,33] The mechanical properties of the as-prepared electrolytes are the key to suppressing the growth of Li dendrites during cycling. The stress-strain curves of the as-prepared electrolytes are plotted in Figure 2c. The MOF-added electrolyte (M-PEO 20 ) has a higher tensile strength than PEO 20 , while the PEO-based electrolyte with C-S-C groups linkage (P(EO 20 -ES)) has a higher extensibility than PEO 20 , indicating that the MOFs can enhance the tensile strength and that C-S-C groups can increase the extensibility of the electrolyte. Notably, the double crosslinked M-P(EO 20 -ES) demonstrates a tensile strength of 2.1 MPa and extensibility of 34% due to the synergistic effect of MOFs and C-S-C groups.
To highlight the superiority of M-P(EO 20 -ES) in ionic conductivity, the Arrhenius plots for the ionic conductivities of SPEs at different temperatures (from 20 to 80°C) are presented, and their activation energy (E a ) for ionic transport was calculated by the Arrhenius formula ( Figure 2d). Since the incorporation of alternate C-S-C groups in the polymer backbone of PEO 20 helps to weaken the coordination between Li þ and the polymer backbone, E a decreases from 0.59 eV (PEO 20 ) to 0.47 eV (P(EO 20 -ES)), indicating that Li þ transport in P(EO 20 -ES) is faster than that in PEO 20 . FTIR confirmed the weakened coordination strength between Li þ and the polymer backbone with the incorporation of C-S-C groups ( Figure S5, Supporting Information). The coordinated C-O-C vibration on P(EO 20 -ES) is much weaker than that on PEO 20 , and an obvious free C-O-C vibration appears. [19] As a result, the increased polymer segment relaxation and weakened coordination between Li þ and the polymer backbone synergistically increase the ionic conductivity by 13 times, from 3.1 Â 10 À6 S cm À1 (PEO 20 ) to 4.0 Â 10 À5 S cm À1 (P(EO 20 -ES)) at room temperature. Moreover, the functionalized MOFs promote the dissociation of Li-salt and increase the Li þ concentration by Lewis acid-base interactions between MOFs and anions, [34] contributing to a higher room temperature ion conductivity (5.8 Â 10 À5 S cm À1 ) and a lower E a (0.43 eV) for M-P(EO 20 -ES). The introduction of an appropriate amount of MOFs can improve Li þ transport in the polymer electrolytes. However, the excessive introduction of MOFs in the polymer electrolyte will lead to low ionic conductivity ( Figure S6, Supporting Information), because the agglomeration between MOF particles blocks the ionic transport pathways.
To verify the advantages of crosslinked M-P(EO 20 -ES), the unfunctionalized MIL-101-NH 2 was mixed into the precursor of P(EO 20 -ES) to obtain a physically mixed MOFs/P(EO 20 -ES) electrolyte. The physically mixed MOFs/P(EO 20 -ES) shows sluggish ionic transport and poor mechanical properties compared with chemically crosslinked M-P(EO 20 -ES) ( Figure S7, Supporting Information), owing to the inhomogeneous dispersion and limited compatibility. The Li þ -transference number of SPEs is closely related to concentration polarization and rate performance for lithium metal batteries, and it was calculated through chronoamperometry curves and corresponding impedance spectra ( Figure S8, Supporting Information). The M-P(EO 20 -ES) shows a higher Li þ -transference number of 0.46 (Figure 2e) than P(EO 20 -ES) and PEO 20 because the interactions between MOFs and anions can effectively hinder the transport of TFSI À . [24] In addition to high ionic conductivity and Li þ -transference number, M-P(EO 20 -ES) electrolyte also shows good electrochemical stability. Linear sweep voltammetry (LSV) was carried out to determine the electrochemical stability windows of the as-prepared SPEs (Figure 2f ). The P(EO 20 -ES) started to oxidatively decompose at 4.1 V due to the poor electrochemical stability of C-S bonds. [22] However, the electrochemical stability window of M-P(EO 20 -ES) reaches as high as 4.6 V, benefiting from the hydrogen bond interaction between the C-S-C/C-O-C group of the polymer chains and the -NH-group of the functionalized MIL-101-NH 2 . [34] The interfacial stability between M-P(EO 20 -ES) and the lithium metal electrode was investigated by performing lithium plating and stripping in symmetrical Li||Li cells using M-P(EO 20 -ES), P(EO 20 -ES) and PEO 20 electrolytes. Figure 3a shows the voltage responses of SPEs at 50°C under 0.1 and 0.1 mAh cm À2 . The Li| PEO 20 |Li cell showed a large overpotential, and the overpotential increased rapidly due to the sluggish ion transport and poor electrochemical stability. For the P(EO 20 -ES) electrolyte, the symmetrical Li||Li cell presented stable cycling with a small overpotential until a short circuit occurred after 220 h, suggesting the poor electrochemical stability, and the mechanical strength of P(EO 20 -ES) was not enough to suppress the growth of Li dendrites. In contrast, the M-P(EO 20 -ES) exhibited more stable lithium plating/stripping over 800 h and enabled the symmetrical Li||Li cell to work steadily for over 2400 h at room temperature and 0.05 mA cm À2 (Figure 3b), indicating that the M-P(EO 20 -ES) electrolyte has superior interfacial stability with lithium metal. To obtain a better understanding of the interfacial stability between M-P(EO 20 -ES) and the lithium metal electrode, the lithium Coulombic efficiency and electrodeposited morphology were investigated by Li||Cu cells. The Coulombic efficiency in the PEO 20 was just only 24% and quickly died out ( Figure S9, Supporting Information), while M-P(EO 20 -ES) displayed a stable lithium Coulombic efficiency of 83% after the initial activation, which significantly surpassed that of P(EO 20 -ES). The more efficient lithium plating/stripping of M-P(EO 20 -ES) can be ascribed to the homogenization of Li þ deposition by the chemically crosslinked MOFs in M-P(EO 20 -ES). [35] To verify this viewpoint, the morphology of the lithium electrodeposition was investigated to intuitively demonstrate the optimization of Li þ deposition by chemically crosslinked MOFs. As shown in Figure 3c, the lithium deposition in the PEO 20 showed an uneven and corroded morphology, indicating a nonuniform Li þ flux and obvious side reactions. Compared with the uneven spherical morphology in P(EO 20 -ES) (Figure 3d), the lithium deposition in M-P(EO 20 -ES) exhibited a smooth and uniform spherical morphology (Figure 3e), corresponding to a uniform Li þ flux. In addition, the conventional commercial electrolyte (1M LiPF 6 in EC/ DEC) was used for comparison ( Figure S10, Supporting Information). Different from the uniform spherical morphology of M-P(EO 20 -ES), the conventional commercial electrolyte showed an uneven dendritic lithium morphology, which is adverse for the stable cycling and safety of lithium metal batteries. [36] Moreover, compared with organic-dominated SEI in the conventional commercial electrolyte, M-P(EO 20 -ES) possesses an inorganic-dominated SEI enriched in LiF, Li 2 S, and Li 2 O, which is believed to further protect the lithium anode and provide fast interfacial ion transport. [37,38] With high ionic conductivity, high Li þ -transference number, wide electrochemical stability window, and good electrochemical stability against the lithium metal, the feasibility of M-P(EO 20     In addition to the intrinsic properties of electrolytes, the interfacial contact between electrodes and electrolytes is also critical for the electrochemical performance of SSLBs. Due to the rigid contact between the SPE and cathode, the transport pathways of Li þ at the SPE/cathode interfaces and inside the porous cathode are inhomogeneous and discontinuous, resulting in large interfacial resistance and low active material utilization, [39,40] especially in the case of high cathode loading. Consequently, the performance of SPEs in most reports was evaluated by low cathode loading (typically <3 mg cm À2 ), which is far from the intention of high energy density. [41,42] To improve the interfacial contact between the cathode and SPE, a cathode-supported SPE strategy is provided as follows (Figure 5a   investigate the improvement of the SPE/cathode interface contact of cathode-supported M-P(EO 20 -ES), cross-sectional scanning electron microscope (SEM) and energy-dispersive spectroscopy (EDS) were conducted (Figure 5b). The cathodesupported M-P(EO 20 -ES) showed an intimate SPE/cathode interface compared with the insufficient interfacial contact in the conventional sandwich configuration ( Figure S15b, Supporting Information). In addition, corresponding EDS mapping revealed that the Cl signal of M-P(EO 20 -ES) was distributed inside the cathode, which confirmed the partial infiltration of M-P(EO 20 -ES) within the porous cathode. The aforementioned result indicates that the cathode-supported M-P(EO 20 -ES) is an effective strategy to construct continuous ionic transport pathways between the electrolyte and cathode.
To show the superiority of this strategy, two types of SSLBs were assembled with lithium metal electrodes and high LiFePO 4 loading (10.0 mg cm À2 ) to evaluate their electrochemical performance at 50°C: 1) a cell assembled with conventional freestanding M-P(EO 20 -ES) (conventional SSLB) and 2) a cell assembled with cathode-supported M-P(EO 20 -ES) (cathode-supported SSLB). As shown in Figure 5c, the interfacial impedance of the cathode-supported SSLB (49 Ω) is much lower than that of the conventional SSLB (180 Ω), corresponding to faster kinetics of the electrochemical reaction. At an increased LiFePO 4 loading of 10.0 mg cm À2 , the conventional SSLB only delivered a low areal capacity of 0.32 mAh cm À2 with a larger overpotential (Figure 5d,e). In contrast, the cathode-supported SSLB enabled a reversible capacity of 143 mAh g À1 with flat charge-discharge plateaus, corresponding to an areal capacity of 1.43 mAh cm À2 . The significant promotion in electrochemical performance is attributed to the intimate interface and developed ionic transport pathways between the electrolyte and cathode. Benefiting from the high active material utilization of 84% at a high LiFePO 4 loading, the cathode-supported SSLB achieves a higher gravimetric energy density of 230 Wh kg À1 compared with previously reported SSLBs (Figure 5f, and Table S2, Supporting Information), demonstrating a promising strategy to boost high energy density SSLBs. [43,44] In addition, the pouch cell was assembled to evaluate the safety property under harsh conditions as shown in Figure 5g. The device can light up an LED device even if the pouch cell is cut in half, suggesting its high safety.

Conclusion
In this work, a double crosslinked PEO-based electrolyte by C-S-C groups and MOFs was proposed, which realized the improved coordination environment and high stability. The incorporation of alternate C-S-C groups in the backbone of PEO facilitates segmental relaxation of the polymer backbone and weakens the coordination between Li þ and the polymer backbone. The double crosslinked structure covalently linked with the MOFs shows an enhanced Li þ -transference number and high electrochemical/mechanical/thermal stability. Therefore, the LiFePO 4 ||Li cells based on M-P(EO 20 -ES) electrolytes showed excellent cycling performance over a wide temperature range (25-100°C). Furthermore, cathode-supported M-P(EO 20 -ES) was fabricated to construct continuous ion transport pathways at the electrolyte/cathode interface, which can reach 84% of the theoretical capacity for the LiFePO 4 electrode with a cathode loading of 10.0 mg cm À2 . This work shows great potential in the rational design of solid-state electrolytes and electrodes for high-safety and high-energy SSLBs.

Experimental Section
Preparation of Functionalized MIL-101-NH 2 : 1) Samples of MIL-101-NH 2 were prepared according to previously reported procedures. [45] Briefly, Cr(NO 3 ) 3 ·9H 2 O (400 mg, 2 mmol), 2-aminoterephthalic acid (180 mg, 2 mmol), and NaOH (100 mg, 5 mmol) were dissolved in deionized water (7.5 mL), and then the mixture solution was transferred to a 25 mL Teflon-lined stainless steel autoclave and kept at 150°C for 12 h. The obtained powders were isolated by centrifugation and washed with dimethylformamide and alcohol. 2) Postsynthetic modification of MIL-101-NH 2 : the as-synthesized MIL-101-NH 2 was mixed with methacrylic anhydride in CH 2 Cl 2 and stirred for 72 h at room temperature. [25] The nanoparticles were washed several times with fresh CH 2 Cl 2 and then dried under vacuum at 40 ºC for 6 h.
Preparation of M-P(EO 20 -ES) Electrolyte: First, 0.4 g PEGDA (M w ¼ 1000), 0.3 g SH-PEG-SH (M w ¼ 1000), and LiTFSI (EO: Li þ ¼20) were dissolved in 500 μL CH 2 Cl 2 solvent. Then functionalized MIL-101-NH 2 (3 wt% of the total weight of PEGDA and SH-PEG-SH) and DMPA (1 wt% of the total weight of PEGDA and SH-PEG-SH) were dispersed ultrasonically in the aforementioned solution for 1 h. The homogeneous suspension was transferred to a glass plate for 10 min to evaporate the CH 2 Cl 2 solvent and exposed to a UV lamp two times every 20 min in a glove box. M-P(EO 20 -ES) was obtained after drying in a vacuum oven at 50°C for 4 h to remove the remaining CH 2 Cl 2 solvent.
Preparation of Cathode-Supported M-P(EO 20 -ES): First, the M-P(EO 20 -ES) precursor solution was directly drop-coated on the as-prepared cathode, and then kept at 60°C for 1 h to allow the CH 2 Cl 2 to permeate completely. Second, the cathode-supported M-P(EO 20 -ES) was cured rapidly under a UV lamp. To ensure a high conversion, the cathode-supported M-P(EO 20 -ES) was polymerized under a UV lamp two times every 20 min. Finally, the cathode-supported M-P(EO 20 -ES) was obtained after drying in a vacuum oven at 50°C for 4 h to remove the remaining CH 2 Cl 2 solvent.
Preparation of the Cathode: 70 wt% LiFePO 4 , 10 wt% Super-P, 10 wt% PVDF, and 10 wt% Homo-SPE (solid polymer electrolyte) [16] were mixed together with NMP solvent to form a uniform slurry and then cast on carbon-coated aluminum foils and dried in a vacuum oven at 60°C for 24 h. The foils were cut into circular cathodes with a diameter of 10 mm, and the LiFePO 4 loading was around 1 mg cm À2 for the cell test. The NCM622 cathode was prepared by a similar procedure. The preparation process of the cathode with high cathode loading was the same as the aforementioned except that the LiFePO 4 loading was %10 mg cm À2 and the proportions are the following: 76 wt% LiFePO 4 , 10 wt% Super-P, and 14 wt% Homo-SPE.
Cell Assembly: The SSLB was prepared using SPE, cathode, and lithium metal as the anode and sealed in a 2025-type cell inside an Ar-filled glovebox. For cathode-supported SSLB, cathode-supported M-P(EO 20 -ES) was used to replace the SPE and LiFePO 4 cathode. The thickness of SPE films is approximately 150 μm, and the thickness of the lithium metal electrodes is %500 μm. For cells with liquid electrolyte, Celgard 2400 and 1 M LiPF 6 in EC/DEC were used to replace the solid polymer electrolytes.
Electrochemical Characterization: Electrochemical properties of the prepared SPEs were determined on Biologic VSP-300 electrochemical workstation. Galvanostatic charge-discharge tests were carried out by a LAND CT2001A instrument. Ionic conductivity measurements were carried out by two blocking electrodes, and the EIS test was scanned in the frequency range of 0.1 Hz to 1 M Hz with a perturbation voltage of 10 mV. The ionic conductivity σ was calculated from the following equation where L (cm) is the thickness of the SPEs, S (cm 2 ) is the contact area between the electrodes, and R is the bulk resistance of the electrolyte. The electrochemical stability windows of the electrolytes were obtained at 50°C by LSV. The tested cells were assembled using stainless steel as the working electrode, and lithium metal as the counter electrode with 2 mV s À1 . The Li þ -transference number of the electrolytes was measured by a chronoamperometry test on a symmetric Li||Li cell with an applied voltage of 10 mV at 50°C. The t Liþ was calculated from the following equation where ΔV is the applied voltage, I 0 and I s are the initial and steady-state currents, respectively, and R 0 and R s are the initial and steady-state resistances obtained by EIS. Materials Characterization: XRD patterns were obtained on a Rigaku diffractometer using Cu Kα radiation. The morphology of the samples was characterized by an SEM (Verios G4 UC, 0.6 nm@15 kV) equipped with an Oxford EDS analysis system (Ultim Max 100). For the EDS mappings, LiClO 4 was used instead of LiTFSI in M-P(EO 20 -ES) to avoid the signal interference of S element in the cathode. TGA was performed using NETZSCH STA 449C with a heating rate of 10°C min À1 from room temperature to 800°C in argon. DSC was performed using a DSC214 NETZSCH with a heating rate of 10°C min À1 from À80 to 100°C in argon. XPS analysis was performed using an ESCALAB 250 instrument with Al Kα radiation (15 kV, 150 W) under the pressure of 4 Â 10 À8 Pa. Raman spectra were obtained using a Wintec alpha300R with 633 nm laser. FTIR spectra were obtained using a Nicolet iS5 iD7 spectrometer. The mechanical properties of SPEs were determined by a dynamic mechanical analyzer (DMA Q800, TA Instrument).

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.